Interaction of ataxin-2 and ataxin-3 in Machado-Joseph disease

Sara Matias Carmo

Silva

Interaction of ataxin-2 and ataxin-3 in

Machado-Joseph disease

Interação da ataxina-2 e ataxina-3 na doença de

Machado-Joseph

2011

Sara Matias Carmo

Silva

Interaction of ataxin-2 and ataxin-3 in

Machado-Joseph disease

Interação da ataxina-2 e ataxina-3 na doença de

Machado-Joseph

Dissertação apresentada à Universidade de Aveiro para cumprimento dos requisitos necessários à obtenção do grau de Mestre em Biologia Molecular e Celular, realizada sob a orientação científica do Professor Doutor Luís Pereira de Almeida do Centro de Neurociências e Biologia Celular de Coimbra e Professor Auxiliar na Faculdade de Farmácia da Universidade de Coimbra da Universidade de Coimbra e de Professora Doutora Margarida da Cruz Fardilha, Professora Auxiliar Convidada da Secção Autónoma de Ciências da Saúde da Universidade de Aveiro.

o júri

presidente

Prof. Doutora Odete Abreu Beirão da Cruz e Silva

Professora Auxiliar com Agregação da Secção Autónoma de Ciências da Saúde - Universidade de Aveiro

Prof. Doutora Ana Cristina Carvalho Rego

Professora Auxiliar da Faculdade de Medicina da Universidade de Coimbra; Centro de Neurociências e Biologia Celular de Coimbra

Prof. Doutor Luís Pereira de Almeida

Professor Auxiliar da Faculdade de Farmácia da Universidade de Coimbra; Centro de Neurociências e Biologia Celular de Coimbra

Prof. Doutora Margarida Sâncio da Cruz Fardilha

Professora Auxiliar Convidada da Secção Autónoma de Ciências da Saúde- Universidade de Aveiro

agradecimentos Ao Professor Doutor Luís Pereira de Almeida agradeço a confiança depositada em mim e a oportunidade que me proporcionou de integrar o seu grupo de trabalho. Agradeço a qualidade da orientação científica bem como toda a disponibilidade e apoio.

Ao Professor Doutor Clévio Nobrega por toda a disponibilidade, compreensão e acima de tudo paciência que sempre demonstrou para comigo. Pela qualidade da orientação do meu trabalho e pelo apoio que nunca faltou. Pela motivação e por tudo o que me ensinou ao longo deste percurso. Enfim, por toda a ajuda, que foi determinante para a conclusão desta dissertação. O meu muito obrigada.

À Professora Doutora Margarida Fardilha pela co-orientação científica deste trabalho, pela disponibilidade e incentivo.

A todos os meus colegas do Grupo de Vectores eTerapia Génica do Centro de Neurociências e Biologia Celular, pela contribuição científica, pela ajuda prestada sempre que me senti perdida e por todo apoio.

À Ana Cristina um obrigado especial, por toda a ajuda que me deu no laboratório, mas principalmente pela amizade, pelos cafés e por mesmo sem saber, me ter ajudado a manter a sanidade mental. Obrigada.

Às minhas amigas de sempre, Catarina, Rita e Sílvia, pela força que longe ou perto sempre me deram, por serem o meu rochedo faz tanto tempo. Por os telefonemas em desespero que sempre tiveram respostas, pelos cafés, pela companhia, por me acolherem na minha volta a Coimbra. Obrigada.

Ao Ricardo, pelos sermões, por todas as vezes que me ralhou para não perder a concentração, mas que nunca me falhou. Obrigada por estares sempre aqui. À minha Avó, que sempre torceu por mim e que mesmo na ausência permanecerá sempre no meu coração.

Ao meu Irmão, cujas palavras duras sempre foram de incentivo, que sempre me incentivaram a ser mais e melhor e porque sem ele o meu caminho seria muito diferente. Mas principalmente por me ter dado o meu maior incentivo, o Afonso.

Aos meus pais, que sempre me deram tudo o que podiam e não podiam, o possível e o impossível e continuam hoje a fazê-lo como no primeiro dia. Obrigada por tudo o que me dão.

A todos aqueles que não referi, mas que não foram esquecidos, aqueles que sempre acreditaram em mim e sempre me apoiaram. Um muito obrigada.

palavras-chave DMJ, ataxina-3, ataxina-2, PABP, PAM2, A2BP1,efeitos modulatórios, interação proteica, CAG, SCA3, SCA2, SCA1, ALS.

resumo A doença de Machado-Joseph também conhecida como ataxia espinhocerebelosa tipo 3 (MJD/SCA3) é uma doença neurodegenerativa hereditária descrita pela primeira vez em emigrantes portugueses da região dos Açores. É a ataxia autossómica dominante mais comum e apresenta manifestações clínicas severas com um desfecho fatal. Esta doença é causada por uma expansão instável do trinucleótido CAG na região codificante do gene MJD1, que se traduz numa cadeia com um número excessivo de glutaminas na ataxina-3. A ataxina-3 mutante adquire propriedades tóxicas e desencadeia uma patogénese complexa com manifestações heterogéneas.

Apesar da expansão poliglutamínica na proteína ataxina-3 ser suficiente para induzir a patogénese da doença e do número de repetições se correlacionar inversamente com a idade de início dos sintomas, não permite explicar toda a diversidade na idade de iníco e nas manifestações clínicas. Há evidências de que a interação entre diferentes proteinas pode afectar quer a idade de ínicio dos sintomas quer a progressão da doença.

Na procura de modificadores da neurodegeneração da DMJ estudos recentes sugerem que a ataxina-2, uma proteína envolvida na ataxia espinocerebelosa do tipo 2 (SCA2) que interage com a proteína ligante da cadeia de poliadenina do RNA mensageiro (PABP) e poderá potenciar a neurodegeneração na DMJ. O objectivo principal deste trabalho foi investigar a interacção entre a ataxina-2, PABP e a ataxina-3 na DMJ.

Utilizando material humano (tecido cerebral e fibroblastos), animais trangénicos para a doença, um modelo lentiviral em roedor e ainda um modelo celular da doença, observámos que os níveis de ataxina-2 e PABP sofrem alterações na DMJ. A imunoreactividade para a ataxina-2 encontra-se aumentada no tecido humano dos doentes, o que também ocorreu também no murganos trangénicos. No modelo lentiviral observámos um recrutamento de ambas as proteínas para o núcleo da célula coincidente com a progressão da doença.

O modelo celular revelou um feedback positivo entre os níveis de ataxina-2 e PABP, que poderá estar na base da sua capacidade de reduzir os níveis de ataxina-3 mutante na doença. Provando assim que os níveis das três proteínas parecem ser interdependentes, permanecendo a questão se esta interação ocorre de forma directa ou indirecta.

Estes estudos sugerem que os níveis e consequente toxicidade de uma proteína poliglutamínica mutada pode ser modulada pela actividade normal de outra. Esta relação sinérgica pode ser crítica para a patogénese da doença e desta forma constituir um potencial alvo terapêutico.

keywords MJD, ataxin-3, ataxin-2, PABP, PAM2, A2BP1,modulatory effect,proteic interactions, CAG, SCA3, SCA2, SCA1, ALS

abstract Machado-Joseph disease also known as Spinocerebellar ataxia type 3 (MJD/SCA3) is a hereditary neurodegenerative disease first described among Portuguese immigrants from Azores. Is the most common autossomal domi-nant spinocerebellar ataxia worldwide and it presents severe clinical features with a fatal outcome. This disease is caused by an unstable expansion of the trinucleotide CAG in the coding region of the MJD1 gene, the mutation is trans-lated into an expanded polyglutamine tract in the protein ataxin-3. The mutant ataxin-3 acquires toxic properties and leads to a downstream of pathogenic mechanisms that make MJD a very heterogeneous disease.

Although the polyglutamine expansion in ataxin-3 is sufficient for the pathogen-esis of the disease and for the number of repeats is correlated inversely with the age of onset, clinical observations suggest that the polyQ expansion does not account for all the diversity in the age of onset and clinical manifestations. Interactions between different proteins may affect the age of onset as well as the disease progression.

In a screen for modifiers of neurodegeneration in MJD, previous studies found ataxin-2, a protein involved in the spinocerebellar ataxia type 2 (SCA2) which interacts with poly (A)-binding protein (PABP) and may potentiate the neuro-degeneration in MJD. The main goal of our research was to investigate the relation between ataxin-2, PABP and ataxin-3 in MJD.

Recurring to human material (brain tissue and fibroblasts), transgenic mice for MJD, a lentiviral-based MJD mouse model and a cellular model of the disease, we observed that levels of ataxin-2 and PABP suffer alterations in MJD. The ataxin-2 immunoreactivity is increased in the human tissue of MJD patients, which also occurred in the transgenic rodents. In the lentiviral model we ob-served both proteins’ recruitment into the nucleus with the progression of the disease.

The cellular model revealed a positive feedback between ataxin-2 and PABP that may to be responsible for their capacity in reducing mutant ataxin-3 levels in the disease. Attesting that the levels of the three proteins seem to be inter-dependent, remaining only the question whether this interaction occurs direct or indirectly.

These studies indicate that the levels and the consequent toxicity of a polyglu-tamine disease protein can be modulated by the normal activity of another. This synergicall relation may be critical for the pathogenesis of the disease and therefore be a potential therapeutic target.

1. Introduction ... 1

1.1. Disease of unstable triplet repeat expansion ... 2

1.2. Polyglutamine diseases ... 3

1.3. Spinocerebellar ataxias ... 6

1.4. Machado-Joseph disease/ Spinocerebellar ataxia type 3 ... 7

1.4.1. Genetic features ... 7 1.4.2. Clinical features ... 8

1.4.3. Neuropathological features ... 9

1.4.4. MJD1 gene ... 10

1.4.5. Ataxin-3 ... 11

1.4.5.1. Structure ... 11

1.4.5.2. Function ... 13

1.4.6. Nuclear inclusions and pathogenesis mechanisms ... 16

1.4.6.1. Nuclear inclusions ... 16

1.4.6.2. Pathogenesis mechanisms... 18

1.5. Ataxin-3 interactions ... 20

1.5.1. Ataxin-3 interaction with other polyglutamine proteins ... 21

1.6. Ataxin-2 ... 23

1.6.1. Ataxin-2 structure ... 23

1.6.2. Ataxin-2 function ... 24

1.6.2.1. Stress granules (ataxin-2 and PABP co-localization site) ... 25

1.6.2.2. A2BP1... 26

1.6.3. Ataxin-2 localization ... 26

1.7. Ataxin-2 proteic interactions ... 27

1.7.1. Ataxin-2 mediates ataxin-1 induced neurodegeneration ... 27

1.7.2. Ataxin-2 intermediate length polyglutamine expansions are a risk factor for

ALS ... 28

1.8. Ataxin-2 connection to PABP determines its functions?...29

1.8.1. PABP ... 31

2.1. Tissue preparation ... 39

2.1.1. Human ... 39

2.1.2. Transgenic and wild type mice ... 39

2.1.1. Lentiviral-based MJD rat model ... 40

2.2. Cell cultures ... 41

2.2.1. Neuroblastoma cell culture ... 41

2.2.2. Human fibroblasts culture ... 41

2.3. Infection /Transfection ... 42

2.3.1. Lentiviral vectors ... 42

2.3.2. N

A infection ... 42

2.3.3. Lipofectamine transfection ... 43

2.4. Immunochemical procedures ... 44

2.4.1. Immunohistochemistry for human brain tissue ... 44

2.4.2. Immunohistochemistry for transgenic and wild-type mouse sections ... 44

2.4.3. Free-floating Immunohistochemistry for lentiviral-based MJD rat model

sections ... 45

2.4.4. Immunohistochemistry quantitative analysis ... 46

2.4.5. Immunocitochemistry of human fibroblasts ... 46

2.4.6. Immunocitochemistry quantitative analysis ... 47

2.5. Protein extraction and Western Blotting ... 47

3. Results ... 48

3.1. Ataxin-2 immunoreactivity is increased in Machado-Joseph disease patient

brain tissue ... 49

3.2. Ataxin-2 and PABP levels are increased in Machado-Joseph disease patient

fibroblasts ... 52

3.3. Ataxin-2 and PABP levels decrease with the progression of the disease in a

lentiviral-based MJD model ... 55

3.4. Ataxin-2 and PABP levels are altered in a MJD transgenic mouse model ... 58

3.5. Ataxin-2 and PABP levels are altered in a cellular model of Machado-Joseph

disease ... 62

3.7. Silencing PABP does not alter ataxin-3 neither ataxin-2 levels but increases

mutant ataxin-3 aggregates ... 69

4. Discussion ... 73

5. Conclusions and Perspectives ... 80

Figure 1. Representation of the possible localization of triplet repeats for various diseases.

(Adapted from Tsuji, 1997) ... 3

Figure 2. Different CAG repeat expansions in Polyglutamine disorders. (Adapted from Schöls

et al., 2004) ... 6

Figure 3. The main areas exhibiting neuronal loss in MJD. Indicated in the figure three of the most affected regions, the cerebellum, the ventral pons and substantia nigra. Large dots indi-cate severe neuronal loss. Blue dots indiindi-cate involvement of extrapyramidal nuclei. Green dots indicate cranial nerve involvement. (Adapted from Taroni and DiDonato, 2004) ... 9

Figure 4. Structure of the human Ataxin-3. Ataxin-3 is mainly composed of a conserved

Josephin domain in the N-terminal side, encoding the ubiquitin protease with the catalytic triad (Cys, His and Asn) and a Nuclear Export Signal (NES). In the flexible C-terminal there are 2 or 3 ubiquitin-interacting motifs (UIM), a Nuclear Localization signal (NLS) and the polyQ tract. ... 12

Figure 5. Inclusion bodies formation. Proteins with an expanded polyQ stretch are prone to

misfold into a β-sheet dominant structure, leading to their assembly into oligomers and amyloid fibrillar aggregates, followed by their accumulation as inclusion bodies within neurons, eventually resulting in neurodegeneration. Dashed arrows indicate structures for which cytotoxicity remains controversial. (Adapted from Popiel et al., 2011) ... 17

Figure 6. Mechanisms of pathogenesis in Machado-Joseph disease. Expanded polyglutamine

proteins might mediate pathogenesis trough a range of common mechanisms; MJD is no exception. The MJD1 gene with expanded CAG repeat is transcribed and translated into ataxin-3 with expanded polyglutamine tract. The expansion results in an aberrant structure of ataxin-3, which forms possible new proteic interactions that wouldn’t happen in normal situation and triggers several events that lead to neurodegeneration in selective areas of the brain. Accumulation of mutant ataxin-3 to form insoluble NIIs recruits components of the UPS and other protein quality-control pathways, furthermore, the normal function of ataxin-3 in the cell can contribute to the impairment of UPS, thus leading to a dysfunction in these quality-control mechanisms. The full-length polyglutamine protein can also be cleaved by proteases to form fragments, which are toxic and might also mediate pathogenic effects. The oligomerization and aggregation and the posttranslational modifications may also contribute to the increase of the neuronal cytotoxicity and dysfunction. The mutant protein affects many other cellular processes, including transcription and RNA metabolism. Other cellular

ubiquitin. (Adapted from Gatchel and Zoghbi, 2005) ... 19

Figure 7. Protein structure of human ataxin-2. The PolyQ tract in the N-terminal region, the

two globular domaisn, Lsm and LsmAD domains, and the PAM2 motif in the C-terminal region. ... 23

Figure 8. Protein structures of human Ataxin-2 and its yeast homologue Pbp1. Both proteins

have a putative RNA binding Lsm domain followed by an as yet uncharacterized Lsm-associated domain LsmAD. ATX2 also contains a PABC-interacting motif named PAM2, which Pbp1 does not contain. (Adapted from Ralser et al., 2005) ... 29

Figure 9. PABP1 structure diagram. The N-terminal consists in the four RRMs, different and

not functionally equivalent. The C-terminal consists in a proline rich region and the PABC. (Adapted from Smith and Gray, 2010) ... 31

Figure 10. Ataxin-2 immunoreactivity is increased in the brain tissue of Machado-Joseph

disease patients. (A,B) Ataxin-2 immunohistochemistry in striatum and cortex of MJD patient’s brains (MJD, n=3), and age-matched controls (Control, n=2). (A) Representative brain sections stained for ataxin-2 to highlight ataxin-2 localization in the striatum and cortex of MJD patients and control (fluorescence microscopy). (B) Quantitative analysis of ataxin-2 immunoreactivity in the striatum and cortex. The immunohistochemistry was performed in triplicate, analyzed in the same day at same exposure. Values are expressed in mean ± SEM. **P<0.01; ***P<0.001 (unpaired Student’s t-test). ... 50

Figure 11. Ataxin-2 is recruited to the nucleus and upregulated in MJD patients fibroblasts

(A,B) Immunocytochemical analysis of fibroblasts of MJD patients (MJD, n=2) and fibroblasts of controls (Control, n=1) by staining nuclei (DAPI, blue), ataxin-3 (Ataxin-3, green) and ataxin-2 (Ataxin-2, red). (A) In MJD, ataxin-2 was recruited to the nucleus, a phenomenon that was not observed in the controls, despite the absence of intranuclear inclusions in MJD patient’s fibroblasts. (B) Immunoreactivity of ataxin-2 in the whole cell was increased in MJD, as well as in the nucleus, comparative to controls. Immunocitochemistry was performed in triplicate; with two slides per group and at least 12 snaps per slide, at the same exposure at the same time. Values are expressed in mean ± SEM. ***P<0.001; ****P<0.0001 (unpaired Student’s t-test). (C, D) Western blotting analysis of lysates of human fibroblasts derived from MJD patients and from a control individual. (C) Representative western-blot probed for endogenous human ataxin-2 (MW: ≅ 148kDa) and tubulin (MW: 55kDa) proteins. (D) Optical

ratio Ataxin-2/ Tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. ***P<0.001 (Student’s t-test). Three replicates of the blots were performed to ensure the reliability of data. ... 52

Figure 12. PABP levels are increased in MJD patient’s fibroblasts. (A,B) Immunocytochemical

analysis of fibroblasts from MJD patients (MJD, n=2) and fibroblasts from control individuals (Control, n=1) with staining for nuclei (DAPI, blue), ataxin-3 (Ataxin-3, green) and PABP (PABP, red). (A) PABP staining is slightly increased in the nucleus of fibroblasts from MJD patients as compared to controls. (B) Immunoreactivity of PABP was significantly higher in MJD, comparative to controls. Immunocitochemistry was performed in triplicate; in two slides per group and at least 12 snaps per slide, at the same exposure in the same day. Values are expressed in mean ± SEM. ****P<0.0001 (unpaired Student’s t-test). (C,D) Western blotting analysis of lysates of fibroblasts derived from MJD patients (MJD) and from a control individual. (C) Representative western-blot probed for endogenous human PABP (MW: 71kDa) and tubulin (MW: 55kDa) proteins. (D) Optical densitometry analysis. An increase in the PABP levels as compared with the control is observed. Each PABP lane was normalized according to the tubulin band. Results were expressed as ratio PABP/ Tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. *P<0.05 (Student’s t-test). Three replicates of the experiment were performed. ... 54

Figure 13. Ataxin-2 levels decrease with the progression of MJD in a lentiviral-based model.

(A,B) Immunohistochemical analysis of mouse sections expressing human wild type and mutant ataxin-3 upon lentiviral injection. The analyses were performed at 2 weeks (n=3, upper lane), 4 weeks (n=3, middle lane) and 8 weeks (n=3, lower lane) post-injection by staining for the nuclei (DAPI, blue), for ataxin-3 (ataxin-3, green) and for ataxin-2 (ataxin-2, red). (A) Representative brain sections stained for ataxin-2 (confocal microscopy), it was notable the decrease of ataxin-2 levels with the progression of time. (B) The immunoreactivity decreased over time, being the decrease more evident in the interval of 2 to 4 weeks. The immunohistochemical analysis was performed in triplicate, with at least three sections per animal, at the same color of staining, analyzed in the same day at same exposure. Values expressed in mean ± SEM. *P<0.05, **P<0.01; ***P<0.001 (One-way Anova). 56

Figure 14. PABP levels decrease with the progression of MJD in a lentiviral model. (A,B)

Immunohistochemical analysis of rat sections expressing human wild type and mutant ataxin-3 upon lentiviral injection. The analyses were performed at 2 weeks (n=3, upper lane),

Representative brain sections stained for PABP and ataxin-3 to highlight PABP levels in the brain sections, it was notable the diminishing of PABP levels with the progression in time (confocal microscopy). (B) The immunoreactivity analysis leads to the conclusion that PABP levels are significantly decreasing over time, being the decrease more accentuated in the interval of 2 to 4 weeks. The immunohistochemmical was performed in triplicate, with at least three sections per animal, at the same color of staining, analyzed in the same day at same exposure. Values expressed in mean ± SEM. *P<0.05, **P<0.01; ***P<0.001 (One-way Anova). ... 57

Figure 15. Ataxin-2 levels are altered in a MJD transgenic mouse model. (A,B)

Immunohistochemical analysis of brain sections of MJD-Q69 mice (n=6) and WT C57BL6 (n=6) for ataxin-2. (A) Ataxin-2 immunoreactivity was observed in sections of WT mice, mainly in the Purkinje cells, and was reduced in the MJD-Q69 mice at this level. (B) Immunoreactivity of ataxin-2 is significantly lower in the MJD Q69 mice, comparative the controls. Immunohistochemical fluorescence analysis was performed in triplicate; with at least 3 snaps per section, at the same exposure in the same day. Values expressed in mean ± SEM. *P<0.05 (unpaired Student’s t-test). (C, D) Western blotting analysis of cerebellar punches of WT C57BL6 (n=3) and MJD Q69 (n=3) mice. (C) Representative western-blot probed for endogenous human ataxin-2 (MW: ≅ 148kDa) and tubulin (MW: 55kDa) proteins. (D) Optical densitometric analysis reveals an increase in the ataxin-2 levels in MDJ-Q69 punches. Each ataxin-2 lane was normalized according to the tubulin band, used as a loading control. Results are expressed as ratio ataxin-2/ Tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. **P<0.01; *** P<0.001 (t-test). Three replicates of the protocol were performed. (E) Representative western blot of punches of striatum, cortex, hippocampus and cerebellum of WT C57BL6 (n=3) and MJD Q69 (n=3) mice, probed for endogenous human ataxin-2 (MW: ≅ 148kDa) and tubulin (MW: 55kDa) proteins. Decreased levels of ataxin-2 were found in the striatum. ... 59

Figure 16. PABP levels are altered in a MJD transgenic mouse model. (A,B)

Immunohistochemical analysis of brain sections of MJD-Q69 mice (n=5) and WT C57BL6 (n=5) mice for PABP. (A) PABP immunohistochemical staining was observed in the sections of WT mice, particularly in the Purkinje cell layer but was significantly decreased in the MJD Q69 brain sections. (B) Immunoreactivity of PABP is significantly lower in the MJD-Q69 mice, compared to the controls. Immunohistochemistry was performed in triplicate; with at least 3

punches of WT C57BL6 (n=3) and MJD Q69 (n=3) mice. (C) Representative western-blot probed for endogenous PABP (MW: 71kDa) and tubulin (MW: 55kDa) proteins. (D) Optical densitometric analysis. A non-significant decrease in the PABP levels was observed in MDJ-Q69 punches. Each PABP lane was normalized according to the tubulin band used as a loading control. Results were expressed as ratio PABP/ Tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. ns.P>0.05 (t-test). Three replicates of the protocol were performed. (E) Representative western blot of punches of striatum, cortex, hippocampus and cerebellum of WT C57BL6 (n=3) and MJD Q69 (n=3) mice, probed for endogenous PABP (MW: 71kDa) and tubulin (MW: 55kDa) proteins. ... 61

Figure 17. Ataxin-2 levels are decreased in a MJD cellular model. (A,B) Western blotting

analysis of non-infected (NI, n=3), expressing human wild-type ataxin-3 (Atx3WT, n=3) and human expanded ataxin-3 (Atx3MUT, n=3) cell lysates. (A) Representative western-blot probed for ataxin-2, ataxin-3 and tubulin. Note the presence of endogenous ataxin-2 (MW: ≅ 148kDa), human Atx3 WT/ATx3-27Q (MW: 50kDa), Atx3Mut/Atx3-72Q (MW: 67kDa) and tubulin (MW: 55kDa). (B) Optical densitometry analysis. Significant decrease in the levels of ataxin-2 was observed in the cells expressing Atx3MUT relatively to the NI cells and the Atx3WT expressing cells. Each ataxin-2 lane was normalized according to the tubulin band. Tubulin was used as a loading control. Results were expressed as ratio Ataxin-2/ tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. **P<0.01; ***P<0.001 (One-Way Anova test). Three replicates of the protocol were performed. ... 63

Figure 18. PABP levels are increased in a MJD cellular model. (A, B) Western blotting analysis

of neuro2A cells: non-infected (NI, n=3), expressing human wild-type ataxin-3 (Atx3WT, n=3) and expressing human expanded ataxin-3 (Atx3MUT, n=3) cell lysates. (A) Representative western-blot probed for PABP, ataxin-3 and tubulin. Note the presence of endogenous PABP (MW: 71kDa), human Atx3 WT/ATx3-27Q (MW: 50kDa), Atx3Mut/Atx3-72Q (MW: 67kDa) and tubulin (MW: 55kDa). (B) Optical densitometry analysis. A significant increase in the levels of PABP was observed in cells expressing Atx3WT and in those expressing Atx3MUT relatively to the NI cells. Each PABP lane was normalized according to the tubulin band. Tubulin was used as a loading control. Results were expressed as ratio PABP/ Tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. *P<0.05; **P<0.01 (One-Way Anova test). It was performed three replicates of the protocol. ... 64

ataxin-3 (Atx3WT, n=3), expressing human wild-type ataxin 3 infected with ataxin 2 (Atx3WT+Atx2, n=3), expressing human expanded ataxin-3 (Atx3MUT, n=3), expressing human expanded ataxin-3 infected with ataxin-2 (Atx3MUT+Atx2, n=3) cells. (A) Representative western-blot probed for Ataxin-2, PABP, ataxin-3 and tubulin. Note the presence of endogenous Ataxin-2 (MW: ≅ 148kDa), PABP (MW: 71kDa), human Atx3 WT/ATx3-27Q (MW: 50kDa), Atx3Mut/Atx3-72Q (MW: 67kDa) and tubulin (MW: 55kDa). (B) Optical densitometry analysis. Significant decrease in the ataxin-3 levels was observed (relative to tubulin and to ataxin-2) in Atx3WT+Atx2 cells relatively to the Atx3WT cells. A similar occurrence was observed in the levels of ataxin-3 with Atx3MUT+Atx2 relatively to non-infected Atx3MUT cells. The ataxin-3 levels were normalized according to the amount loaded (tubulin band). Results are expressed as ratio ataxin-3/tubulin and ataxin-3/ataxin-2 (n=3 for each experimental set). Values are expressed as mean ± SEM. **P<0.01; ***P<0.001 (Student’s t-test). Three replicates of the protocol were performed. (C) Optical densitometry analysis. Significant increase in PABP levels was detected (relative to tubulin) in Atx3WT+Atx2 cells relatively to the Atx2WT cells. A similar occurrence was observed in the levels of PABP in Atx3MUT+Atx2 relatively to Atx3MUt cells. Each PABP lane was normalized according to the tubulin band. Tubulin was used as a loading control. Results were expressed as ratio PABP/ Tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. *P<0.05; ***P<0.001 (Student’s t-test). ... 66

Figure 20. PABP overexpression decreases the levels of ataxin-3 expression and increases

the levels of ataxin-2 expression. (A-C) Western blotting analysis of non-infected cells (NI, n=3), or cells infected with lentiviral vectors encoding, respectively: human wild-type ataxin-3 (Atxataxin-3WT, n=ataxin-3), human wild-type ataxin ataxin-3 and PABP (Atxataxin-3WT+PABP, n=ataxin-3), human expanded ataxin-3 (Atx3MUT, n=3), human expanded ataxin-3 and PABP (Atx3MUT+PABP, n=3) cells. (A) Representative western-blot probed for ataxin-2, PABP, ataxin-3 and tubulin. Note the presence of endogenous ataxin-2 (MW: ≅ 148kDa), PABP (MW: 71kDa), human Atx3 WT/ATx3-27Q (MW: 50kDa), Atx3Mut/Atx3-72Q (MW: 67kDa) and tubulin (MW: 55kDa). (B) Optical densitometry analysis. Significant decrease in ataxin-3 levels (relative to tubulin and to PABP) was observed in Atx3WT+PABP cells relatively to the Atx3WT cells. Similarly, the levels of ataxin-3 decreased with Atx3MUT+PABP relatively to non-infected Atx3MUt cells. Ataxin-3 levels were normalized according to the amount loaded (tubulin band). Results are expressed as ratio ataxin-3/tubulin and ataxin-3/PABP (n=3 for each experimental set).

were no significant differences in Atx3WT+PABP and Atx3WT cells. Significant increase in ataxin-2 levels (relative to tubulin) in Atx3MUT+Atx2 relatively Atx3MUT cells. Each Ataxin-2 lane was normalized according to the tubulin band. Tubulin was used as a loading control. Results were expressed as ratio PABP/ Tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. ns; P>0.05; ***P<0.001 (Student’s t-test). ... 68

Figure 21. PABP silencing increases aggregate formation. (A-E) Cell lysates western blotting

analysis of Atx3MUT cells transfected with shLuc (controls) (n=3) and transfected with shPABP (n=3). (A) Representative western-blot probed for Ataxin-2, PABP, ataxin-3 and tubulin. Note the presence of endogenous ataxin-2 (MW: ≅ 148kDa), PABP (MW: 71kDa), ataxin-3 aggregates (MW> 250kDa), Atx3Mut/Atx3-72Q (MW: 67kDa) and tubulin (MW: 55kDa). (B) Optical densitometry analysis. It was possible to observe a decrease in the levels of PABP (relative to tubulin) in the cells transfected with shPABP relatively the control cells (shLuc), as expected. The PABP levels were normalized according to the amount loaded (tubulin band). Results were expressed as ratio PABP/tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. **P<0.01; **P<0.01 (Student’s t-test). (C) Optical densitometry analysis. There were no significant differences in ataxin-2 between the control cells and the silenced cells. The ataxin-2 levels were normalized according to the amount loaded (tubulin band). Results were expressed as ratio ataxin-2/tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. ns.P>0.05 (Student’s t-test). (D) Optical densitometry analysis. There were no significant differences in ataxin-3 between the control cells and the silenced cells. The ataxin-3 levels were normalized according to the amount loaded (tubulin band). Results were expressed as ratio ataxin-3/tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. ns.P>0.05 (Student’s t-test). (E) Optical densitometry analysis. It was verified an increase in the aggregates of mutant ataxin-3 (found in the stacking) in the silenced cells when compared to the control cells. The aggregates levels were normalized according to the amount loaded (tubulin band). Results were expressed as ratio aggregates/tubulin (n=3 for each experimental set). Values are expressed as mean ± SEM. **P<0.01 (Student’s t-test). ... 70

Table 1. Polyglutamine diseases... 4

Chapter I

Introduction

Page | 1

1.1. Diseases of unstable triplet repeat expansions

An important development in the understanding of the pathogenesis of a group of neurodegenerative diseases occurred more than 20 years ago, with the identification of a mutational mechanism that is behind their pathogenesis, the expansion of unstable nucleotide repeats (Koshy and Zoghbi, 1997; Gatchel and Zoghbi 2005). In the case of triplet repeat diseases, they are caused by an unstable expansion of a trinucleotide repeat (Gatchel

and Zoghbi 2005). In normal conditions, there is a stable transmission of the number of

repeats from the parents, but when there is a mutation involving excessive repetition of the triplet over a certain threshold it becomes unstable in the parental transmission, leading to variations of the number of repeats from one generation to the next. These intergenerational instability leads to an earlier onset and a more severe phenotype in successive generations of an affected family (Timchenco and Caskey, 1996; Tsuji, 1997; Koshy and Zoghbi, 1997).

These diseases are classified depending on the genetic context of the unstable repeat; it may occur an expansion of coding repeats (exons) or an expansion of non-coding repeats (5’UTR, 3’UTR or introns) (Fig.1). The expansion of non-coding repeats, leads to the alteration of gene expression, which reflects in either a loss of protein function or altered RNA function (Tsuji, 1997; Gatchel and Zoghbi, 2005). The other subclass of these disorders is caused by the expansion of an unstable Cytosine-Adenine-Guanine (CAG) trinucleotide repeat within the protein-encoding region, which translates into an abnormally long stretch of glutamine residues, conferring a toxic function to the protein unrelated to its normal function. These diseases are called polyglutamine diseases and are characterized by mutant proteins with expanded polyglutamine (polyQ) tracts (Tsuji, 1997; Ross and Poirer, 2004; Gatchel and

Page | 2

Figure 1:Representation of the possible localization of triplet repeats for various diseases. (Adapted from Tsuji, 1997).

Page | 3

1.2. Polyglutamine diseases

The polyglutamine disorders are autosomal and are dominantly inherited (except SBMA (spinal bulbar muscular atrophy), an X-linked recessive disorder), representing the most common form of inherited neurodegenerative disease. The first polyglutamine disease to be discovered was SBMA in 1991 (La Spada and Keneth Fischbeck, 1991). They are caused by a polyglutamine (polyQ) expansion within their active coding unrelated proteins, including Huntington's disease (HD), dentatorubral-pallidoluysian atrophy (DRPLA), spinal bulbar muscular atrophy (SBMA), and the spinocerebellar ataxias (SCA1, 2, 3, 6, 7 and 17) (Table1) (Orr and Zoghbi, 2000).

Table 1. Polyglutamine diseases

*Main regions affected.

Disease Disease Protein Normal subcellular localization

Affected brain regions*

Huntington’s disease (HD) Huntingtin (htt) Cytoplasmic Striatum and Cortex

Spinal and Bulbar Muscular Atrophy (SBMA)

Androgen receptor (AR)

Nuclear and Cytoplasmic Motor Neurons

Dentatorubal- Pallidoluysian Atrophy (DRPLA)

Atrophin-1 Nuclear and Cytoplasmic Central Cortex

Spinocerebellar ataxia 1 (SCA1)

Ataxin 1 Nuclear and Cytoplasmic Cerebellum

Spinocerebellar ataxia 2 (SCA2)

Ataxin-2 Cytoplasmic Cerebellar Purkinje Cells

Spinocerebellar ataxia 3/ Machado-Joseph disease

(SCA3/MJD)

Ataxin3 Nuclear and Cytoplasmic Cerebellum, Brain Stem, Basal Ganglia, Ventral pons

and Substantia nigra

Spinocerebellar ataxia 6 (SCA6)

Ataxin6 Nuclear and Cytoplasmic Cerebellar Purkinje Cells

Spinocerebellar ataxia 7 (SCA7)

Ataxin7 Nuclear and Cytoplasmic Cerebellar Purkinje Cells, Brain Stem and Spinal Cord

Spinocerebellar ataxia 17 (SCA17)

Page | 4

The polyglutamine expansion diseases have onset in midlife, leading to the death of the patient 10-30 years after the appearance of the first symptoms (Cummings and Zoghbi, 2000;

Orr and Zoghby, 2007; Macedo-Ribeiro et al., 2009). In general the age of onset inversely

correlates with the size of expanded CAG repeats, and there are evidences that point to the interaction of the normal repeat with the expanded CAG repeat influencing the age of onset (Maciel et al., 1995; Maruyama et al., 1995; Gusella and MacDonald, 2000, Djousse et al., 2003).

The genetic similarities among these disorders strongly support the hypothesis that these diseases share a common mechanism of pathogenesis, mainly based in the toxic properties of the polyglutamine tract. The expanded proteins of polyQ disorders have an enhanced ability to form intranuclear aggregates/inclusions (Ross and Poirer 2004), and in some cases also cytoplasmic inclusions in the disease affected regions (Ciechanover and Brundin, 2003; Bennet

et al., 2005). Neuronal nuclear inclusions (NII) are considered to be a histopathological

hall-mark of the polyQ diseases. Although the role of nuclear inclusions in pathology is not com-pletely understood, it is clear that these inclusions result from the nuclear accumulation of polyQ expanded proteins, which are normally in the cytoplasm. Mutant polyQ proteins in the nucleus can abnormally interact with nuclear proteins, such as transcription factors, leading to transcriptional dysregulation (Havel et al., 2009). The presence of these NI of mutant pro-tein focused attention on the nucleus as the subcellular site determinant for the pathogenesis of these diseases, with some suggesting that inclusions were pathogenic (Arrasate et al. 2004) and others questioning this pathogenicity of aggregates/inclusions (Orr and Zoghbi, 2007).

It is becoming more apparent that the polyQ tract by itself; even though toxic and suffi-cient to cause polyQ pathology, does not account for all the symptoms of the polyQ disorders nor for all the similarities observed between them. It is possible that each polyglutamine dis-order may be the result of the expanded polyglutamine tract action in the context of the “host” protein. This means that the normal function and the interactions of each polygluta-mine protein is determinant to define the pathogenic pathway of each disorder, which ex-plains the similarities found between disorders as well as their differences (Orr 2001; Gatchel

Page | 5

1.3. Spinocerebellar ataxias

Autossomal dominant Spinocerebellar Ataxias (SCAs) are a group of neurodegenerative diseases, with some of them being also polyglutamine disorders (SCA1, 2, 3, 6, 7 and SCA17) (Fig. 2). These disorders are clinically and genetically very divergent and are characterized by progressive cerebellar ataxia, which consists in the loss of motor coordination, showing unsteady movements and staggering gait (Zoghbi, 2000). Ataxia results from variable degeneration of neurons in the cerebellar cortex, brain stem, spinocerebellar tracts and their afferent/efferent connections. Although ataxia is the prominent symptom in SCAs, few mutations cause an almost pure cerebellar syndrome isolating neurodegeneration to the cerebellar cortex. On the contrary, most SCAs are multisystemic disorders presenting clinical variability. The ataxia may be accompanied by numerous other symptoms that vary between different SCAs. One common ground between all of these diseases is the progressive neurodegeneration and its fatal outcome (Maruyama et al., 2002; Dueñas et al., 2006).

Figure 2. Different CAG repeat expansions in Polyglutamine disorders. (Adapted from Schöls et al., 2004).

The most striking characteristics in SCAs as unstable triplet expansion diseases are its unstable and heterogeneous expression, observed even in the same family (Moseley, 1998;

Schöls, 2004); and the age of onset of the disease that is inversely correlated with the repeat

length. The age of onset of the clinical symptoms is between 30 and 50 years of age, but there are cases of childhood onset caused by extreme expansions between generations. (Schöls et

Page | 6

1.4. Machado Joseph Disease/Spinocerebellar ataxia

type 3

Machado-Joseph disease (MJD) or spinocerebellar ataxia type 3 (SCA3) is the most frequent autossomal dominantly inherited cerebellar ataxia worldwide, comprising between 15% and 45% of all SCAs (Schöls et al., 1995; Riess et al., 2008). This neurodegenerative disease was first described in 1972 among immigrants native of the Portuguese Azorean Island, São Miguel. Later on, other cases were described in descents of Azorean immigrants from the Flores Island (Nakano et al., 1972; Rosenberg et al., 1976; Rosenberg et al., 1978); and the disorder was subsequently also identified in Brazil, Japan, China, and Australia (Sequeiros

and Coutinho, 1993; Sudarsky and Coutinho, 1995).

MJD is caused by an unstable expansion of a CAG repeat in the coding region of the MJD1 gene mapped on chromosome 14q32.1 (Kawaguchi et al., 1994). The MJD1 gene encodes ataxin-3, a 42 kDa protein that resides in both nucleus and cytoplasm (Paulson et al., 1997a). Ataxin-3 is a polyubiquitin-binding protein whose physiological function is related to ubiquitin-mediated proteolysis (Burnett et al., 2003; Chai et al., 2004). The mutation results in a long polyglutamine chain at the C-terminus of ataxin-3; the CAG repeats range from 10 to 44 in normal population and from 45 to 87 in MJD patients (Cummings and Zoghbi, 2000; Maciel

et al., 2001; Padiath et al., 2005) In MJD, like for other polyQ disorders, an inverse correlation

between the number of CAG repeat and the age of onset of the disease is verified (Maruyama

et al., 1995)

1.4.1. Genetic features

The common underlying genetic basis of MJD is the expansion of a CAG repeat region beyond a certain threshold. There is an overlap of normal repeat sizes (up to 47 repeats) in the CAG tract of the MJD1 gene. As shown in other polyQ diseases such as Huntington’s disease, individuals with a repeat length in the overlapping region do not always manifest the disease in what is designated as incomplete penetrance of the disease (Padiath et al., 2005). This repeat tract is not exclusively a CAG stretch, as variant CAA and AGG triplets were also found in normal alleles (Kawaguchi et al., 1994).

It was described somatic mosaicism of the repeat size in CAG repeat disorders, where different cells of the same individual carry different repeat sizes. This somatic instability has

Page | 7

also been described for MJD in different brain areas, there were found smaller repeat sizes in cerebellar cortex that in other brain regions (Lopes-Cendes et al., 1996; Hashida et al., 1997). Even though the correlation between CAG repeats and the age at onset in MJD is well documented and proved to be a genetic feature of MJD, the length of CAG repeats does not account for all the variation observed in the patients. Also, the variance of the age at onset in MJD patients with similar number of repeats, suggest that additional factors influence the age at onset. It is predicted that a healthy life style leads to a later disease manifestation; it has also been shown that females manifest the disorder later in live as compared to their brothers even with the same expanded repeat size or longer normal CAG repeat sizes (DeStefano et al., 1995 ). CAG repeats are thought to be responsible for 45-60% of variability in age of onset, nevertheless other factors remain to be identified, and there is evidence that similarly to other polyQ diseases, proteic interactions may influence the variability observed in clinical phenotypes and age of onset (Maciel et al., 1995; Van de Warrenburg et al., 2002).

1.4.2. Clinical features

The clinical spectrum of MJD is highly pleomorphic. Based on these clinical manifestations four subphenotypes have been suggested (Coutinho and Andrade, 1978; Schöls et al., 1996;

Riess et al., 2008), which during the progression of the disease may evolve from one type to

other type (Fowler, 1984). Recently it an additional MJD type (V) has been proposed based in a homozygous 33-years old patient of Portuguese/Brazilian descent (Lysenko et al. 2010). Therefore the proposed five clinical subtypes of MJD are:

 Type I: Extrapyramidal (dystonia and bradykinesia)+ Pyramidal deficits + early onset (10-30 years);

 Type II: Pyramidal + cerebellar deficits + intermediate onset (20-50 years);

 Type III: Cerebellar deficits + peripheral neuropathy + late onset (40-75 years);

 Type IV: Neuropathy + Parkinsonism with variable onset;

 Type V: Pronounced Extrapyramidal signs + Pyramidal deficits + spastic paraplegia + variable age of onset (4-43 years).

Other clinical features of MJD include ophtalmoplegia, double vision, faciolingual fasciculation, dysphagia, weight loss without loss of appetite, incontinence, restless legs syndrome, eyelid retraction, postural instability, difficulty with swallowing. Cognitive disturbances are mild and rarely develop to relevant dementia (Sequeiros and Coutinho, 1993;

Sudarsky and Coutinho, 1995; Bürk et al., 2003). Clinically these symptoms may fluctuate in

Page | 8

onset. For example, individuals with early onset of MJD may suffer more from dystonia and bradykinesia than from ataxia (type I). In addition, the latest onset patients, beginning after age 50, may experience as much difficulties resulting from neuropathy as from ataxia (Paulson et al., 2007; Riess et al., 2008).

1.4.3 Neuropathological features

The neuropathological alterations of MJD in the brain consist in neuronal loss in multiple systems. The degeneration is not confined to a single specific region of the brain, but on the contrary is quite widespread which contributes to the multiple clinical impairments observed in the disease. The disorder involves a combination of neurodegeneration in the spinocerebellar tract with other areas of the central nervous system (Fig. 3). The neuropathology involves cerebellar systems (dentate nucleus and pontine nuclei), cranial nerve motor nuclei, substantia nigra, and striatum. A relative preservation of the cerebellar cortex, particularly Purkinje cells at initial stages and inferior olive is reported (Rosenberg,

1992; Coutinho and Andrade, 1995; Durr et al., 1996; Alves et al., 2008a). A marked

degeneration of Clarkes’s column nuclei and vestibular and pontine nuclei is observed (Durr

et al., 1996). Marked neuronal loss is also observed in the anterior horn of the spinal cord,

and motor nuclei of the brainstem (Rub et al., 2008).

Figure 3. The main areas exhibiting neuronal loss in MJD. Indicated in the figure three of the most

affected regions, the cerebellum, the ventral pons and substantia nigra. Large dots indicate severe neu-ronal loss. Blue dots indicate involvement of extrapyramidal nuclei. Green dots indicate cranial nerve involvement. (Adapted from Taroni and DiDonato, 2004)

Page | 9

Studies regarding the localization of mutant ataxin-3 in the brain and its preferential accumulation allowed to describe with much more accuracy the regions affected by the disease itself. Upon macroscopic examination MJD patients brains (disease duration over 15 years) display depigmentation of the substantia nigra, atrophy of the cerebellum as well as atrophy of cranial nerves III to XII. The depigmentation of the substantia nigra is the cause of parkinsonian features, however, not all the patients with degenerated substantia nigra show parkinsonism (Rub et al., 2003; Rub et al., 2006). The majority of these MJD patient’s brains has lower weight than brains from individuals without medical histories of neurological or psychiatric diseases (Iwabuchi et al., 1999). The study of human brains of normal individuals and MJD patients suggests that the cerebellar dentate nucleus and the substantia nigra are among early targets of the degenerative process of MJD, while the thalamus and the cranial nerves (dorsal motor vagal and ambiguous nuclei) are affected only late during the course of the disease (Riess et al., 2008). The involvement of the striatum in the affected brain areas of MJD was recently described (Alves et al., 2008a) and may provide an explanation for symptoms found in some MJD patients such as dystonia and chorea (Lee et al., 2003). Chorea and dystonia have been consistently associated with lesions in the striatum. Chorea involves dysfunction of the indirect pathway from the caudate and putamen to the internal globus

pallidus, whereas dystonia is generated by dysfunction of the direct pathway (Janavs et al., 1998).

One important hallmark of neurodegeneration in the brains of MJD patients is the neuronal intranuclear inclusions (NIIs), which have been observed in patients, transgenic animal models and cellular models (Ikeda et al., 1996; Schmidt et al., 1998; Evert et al., 1999). The NIIs role in the pathogenesis is controversial. Initially associated to the pathogenic mechanism, they may actually contribute to delay the pathogenesis by concealing the toxic mutant ataxin-3 within inert structures reducing its toxicity.

1.4.4. MJD1 gene

The genetic locus for Machado-Joseph disease, MJD1, was identified in 1994 and mapped to chromosome 14q32.1 (Kawaguchi et al., 1994). The MJD1 gene spans about 48Kb with one long open reading frame (17776-base pairs) and is composed of 11 exons. It was identified as a novel gene containing CAG repeats that are located in exon 10 (Ichikawa et al., 2001). The normal MJD1 gene encodes ataxin-3 (ATXN3), a polyubiquitin-binding protein that is mutated in MJD. The mutation results in an expanded polyglutamine tract at the C-terminus of ataxin-3 (Durr et al., 1996). The CAG repeat contains two variant sequences, CAA and AAG, at three

Page | 10

positions that are also translated into a polyglutamine tract at C-terminal region of the protein ataxin-3. The normal allele of MJD1 has usually from 10 to 44 CAG repeats, there is an intermediate state where between 45 and 51 CAG repeats some individuals may manifest the disease and MJD patients alleles range from 52 up to 87 repeats (Cummings and Zoghbi, 2000;

Maciel et al., 2001;Padiath et al., 2005). MJD’s high threshold is a particularity of this disorder

since in other polyglutamine disorders repeats over 36 to 40 become pathogenic (Kawaguchi

et al., 1994; Ichikawa et al., 2001; do Carmo Costa et al., 2004; Rodrigues et al., 2007).

1.4.5. Ataxin-3

Human ataxin-3 is the MJD1 gene product, a ubiquitously expressed protein found in the genome of several species, ranging from nematodes to humans, and plants (Albrecht et al.,

2003). In mice and humans, despite the localized neuronal degeneration observed in MJD

patients, ataxin-3 displays a ubiquitous expression among different body tissues and cell types. MJD1 was found to be widely expressed throughout the brain, though different regions present varying expression levels (Trottier et al., 1998). Although the precise biological func-tion of ataxin-3 still remains poorly understood, evidence supports its participafunc-tion in several pathways related to protein homeostasis maintenance (clearance of misfolded and damaged proteins) (Buchberger, 2002; Doss-Pepe et al., 2003; Burnett et al., 2003; Albrecht et al., 2004;

Chai et al., 2004; Mao et al., 2005; Berke et al., 2005), transcriptional regulation (Evert et al., 2003; Rodrigues et al., 2007), cytoskeleton regulation (Rodrigues et al., 2010) and myogenesis

(do Carmo Costa et al., 2010), cellular activities whose deregulation can compromise cell func-tioning and survival.

1.4.5.1. Structure

The wild type ataxin-3 has an approximated globular weight of 42 KDa in normal individ-uals, possesses 376 amino acids (including 22 glutamines of the polyQ tract) and is widely expressed in the brain and in the body existing in both nucleus and cytoplasm of various cell types (Paulson et al., 1997a; Trottier et al., 1998; Ichikawa et al., 2001). Ataxin-3 consists of a globular deubiquitinating N-terminal Josephin domain (20 KDa), able to resist proteolysis, a flexible protease accessible C-terminal containing two ubiquitin interacting motifs (UIM’s) and a polyQ region of variable length, whose expansion beyond a certain threshold is associ-ated with MJD (Fig.4) (Burnett et al., 2003; Masino et al., 2003; Scheel et al., 2003). The most

Page | 11

common isoform found in the human brain has an extra UIM localized in the C-terminal re-gion, downstream the polyQ sequence (Harris et al., 2010). The two conserved UIMs located N-terminally of the polyQ region are α-helical structures separated by a short flexible linker region and act cooperatively when binding ubiquitin (Ub); the affinity of the two tandem mo-tifs is greater than that of each individual UIM (Song et al., 2010).

Figure 4. Structure of the human Ataxin-3. Ataxin-3 is mainly composed of a conserved Josephin domain in the N-terminal side, encoding the ubiquitin protease with the catalytic triad (Cys, His and Asn) and a Nuclear Export Signal (NES). In the flexible C-terminal there are 2 or 3 ubiquitin-interacting motifs (UIM), a Nuclear Localization signal (NLS) and the polyQ tract.

A highly conserved, putative nuclear localization signal (NLS) is found upstream of the polyQ tract (Tait et al., 1998; Albrecht et al., 2004

)

et al., 2009). Furthermore, two nuclear nuclear export signals (NES) with significant activity

were identified in ataxin-3, following the Josephin domain. (Antony et al., 2009; la Cour et al.,

2003). Even though the NLS has already been identified, the non-expanded protein is present

both in the nucleus and in the cytoplasm; indicating that ataxin-3 possesses nucleocytoplas-mic shuttling activity (Tait et al., 1998; Trottier et al., 1998; Antony et al., 2009). Evidences support the hypothesis that when in a non-aggregated state, the ataxin-3 polyQ is in a ran-dom coil conformation (Chen et al., 2001). Polyglutamine abnormal expansion causes ataxin-3 monomeric peptides to aggregate into β-rich amyloid-like fibrillar structures. The identity of pathogenic species remains unknown – whether the protein alone, the putative intermediates of the fibrillization pathway, or the soluble mature fibrils. Nevertheless, taking in account the similarities with other misfolding amiloidophaties it is possible that protofibrillar intermedi-ates may be far more toxic than large aggregintermedi-ates (Gales et al., 2005).

Page | 12

Wild-type ataxin-3 is evenly distributed throughout the cell, particularly around the nu-clear envelope, whereas mutant ataxin-3 accumulates into intranunu-clear aggregates and is depleted from the cytoplasm. Wild-type ataxin-3 co-localizes to ubiquitin-rich aggregates found in normal aged brains, known as Marinesco bodies (Fujigasaki et al., 2000), is also found in ubiquitinated inclusions of mutant polyQ proteins (Chai et al., 2001) and directly combined with the proteasome (Doss-Pepe et al., 2003).

1.4.5.2. Ataxin-3 function

Ataxin-3 is a de-ubiquitinating (DUB) enzyme present in the nucleus and in the cytoplasm of the cell that appears to function in cellular pathways regulating cellular vigilance, homeo-stasis and transcriptional processes. DUBs are a group of enzymes responsible in the cell for removing ubiquitin (Ub) or polyUb chains from a target protein, processing Ub pre-proteins and remodeling or disassembling bound or unbound polyUb chains (Reyes-Turcu et al., 2009;

Reyes-Turcu and Wilkinson, 2009). Ataxin-3 is an ubiquitin-binding protein capable of binding

and cleaving specific types of large ubiquitin chains, removing ubiquitin from substrates to allow proteosomal degradation. Ataxin-3 binds to ubiquitin from polyubiquitinated sub-strates by the UIM’s, allowing cleavage by ataxin-3 N-terminal Josephin domain. Ubiquitina-tion fulfills many cellular funcUbiquitina-tions in the cytoplasmic trafficking, guiding specific proteins through the endocytic pathways and targeting proteins for the proteasome (Buchberger,

2002; Doss-Pepe et al., 2003; Burnett et al., 2003; Chai et al., 2004; Mao et al., 2005; Berke et al., 2005). In the case of ataxin-3, inhibition of the catalytic activity results in the increase of

polyubiquitinated proteins (localized primarily in the nucleus), to a degree similar to what is observed when the proteasome is inhibited (Berke et al., 2005). This supports the idea of ataxin-3 involvement with the polyubiquitinated proteins targeted for proteasomal degrada-tion, therefore modulating ubiquitin-dependent mechanisms.

These ataxin-3 characteristics suggested a possible role in the Ubiquitin-Proteasome Sys-tem (UPS); the UPS is involved in the processing of mutant or damaged proteins (UPP), DNA repair, chromatin remodeling, cell cycle progression, subcellular localization and pathway signaling (Reyes-Turcu et al., 2009; Reyes-Turcu and Wilkinson, 2009; Weissman, 2001). Its impairment is related to several human diseases, such as polyQ diseases and other neuro-degenerative pathologies. When the enzymatic activity of ataxin-3 is suppressed, it no longer can suppress abnormal protein’s toxicity and becomes somewhat toxic by itself leading to the impairment of the UPS, resulting in cell death (Burnett and Pittman, 2005). Considering its

Page | 13

function it is no surprise that several genes involved in the UPS vigilance pathway were al-tered in the absence of wild-type ataxin-3 or in the presence of mutant ataxin-3 (Rodrigues et

al., 2007). This evidence may be supported by the presence of ubiquitin in ataxin-3

aggre-gates; that may be a consequence of ataxin-3 being itself a substrate for ubiquitination, and thus being degraded by the UPS (Matsumoto et al., 2004; Berke et al., 2005).

Ataxin-3 function may also have relevance in the Endoplasmic Reticulum Associated Pro-tein Degradation (ERAD). There is an ongoing controversia regarding whether ataxin-3 pro-motes or decreases degradation by this pathway (Wang et al., 2006; Zhong and Pittman,

2006); however, unexpectedly recent evidences suggest that ataxin-3 inhibits the ERAD,

which can be explained by proteosomal degradation of ataxin-3 (Riess et al., 2008). Ataxin-3 has been found to associate with the proteasome itself through its N-terminal region (Doss-

Pepe et al., 2003), but this interaction may not be very strong or even direct (Todi et al., 2007).

Functioning with these interactors, ataxin-3 may act in a number of different ways, (a) trim-ming polyUb chains of a substrate, thus facilitating the subsequent disassembly of the chain by proteasome-associated DUBs, (b) editing polyUb chains in order to guarantee that the strate is correctly targeted for degradation, or (c) functioning as a transiently associated sub-unit of the proteasome and recognizing some of its substrates (Boeddrich et al., 2006; Wang et

al., 2008).

Another instance where ataxin-3 has been associated with cellular quality control mech-anisms, is its proposed role in aggresome formation. Aggresomes are misfolded protein ag-gregates that form near the microtubule-organizing center (MTOC) when the proteasome is not able to deal with misfolded proteins. These structures seem to be of physiologic im-portance, since those defective proteins are then degraded by lysosomes, contributing to the maintenance of cellular homeostasis (Markossian and Kurganov, 2004). Endogenous ataxin-3 seems to play a role in the regulation of aggresome formation, having been shown to co-localize with aggresome and preaggresome particles and to be important in the formation of the aggresomes themselves (Burnett and Pittman, 2005). Ataxin-3 also associates with dynein and histone deacetylase 6 (HDAC6), both constituents of the complex responsible for the transport of misfolded proteins to the MTOC (Burnett and Pittman, 2005), and was recently shown to interact with tubulin and microtubule-associated protein 2 (MAP2), two other con-stituents of the cytoskeleton (Mazzucchelli et al., 2009; Rodrigues et al., 2010). Along with its effect in increasing ataxin-3 catalytic activity, ubiquitination of this protein also potentiates its ability to promote aggresome formation (Todi et al., 2010).

Page | 14

Considering the structure of atax3, as well as its localization within the cell and its in-teractions with other proteins, it is possible that ataxin-3 might be involved in transcriptional processes. Due to its characteristics, ataxin-3 has several ways of interacting with the tran-scription. First and indirectly, through its DUB activity, ataxin-3 can modulate the degrada-tion of transcripdegrada-tion factors and repressors by removing the polyubiquitin degradadegrada-tion signal (Li et al., 2002). In addition, ataxin-3 may have a more direct effect over transcription regula-tion since it has a NLS that allows its transportaregula-tion to the nucleus, where it can regulate transcription. Recent studies demonstrated that ataxin-3 is actively transported across the nuclear envelope, being actively shuttled from the cytoplasm to the nucleus and vice versa (Chai et al., 2002; Antony et al., 2009; Macedo-Ribeiro et al., 2009). The nuclear accumulation of ataxin-3 following stress supports the hypothesis of a nuclear function such as transcrip-tion regulatranscrip-tion (Reina et al., 2010). Accordingly, it has also been suggested that ataxin-3 is able to bind histones through its Josephin domain and interact with histone acetyltransferas-es, which work as transcriptional co-activators, thus repressing histone acetylation and tran-scription. Altered protein acetylation has already been implicated in polyQ disease processes (Li et al., 2002; Bodai et al., 2003; Zhang, 2003). The possible role of ataxin-3 in degradation pathways and in transcriptional regulation may be cell-specific, potentially explaining why only a subset of neurons is affected in the human disease despite the widespread expression of normal and mutant ataxin-3 (Rodrigues et al., 2007).

The importance of ataxin-3 interaction with components of the cytoskeleton such as tu-bulin, MAP2 and dynein may not be limited to its possible role in aggresome formation

(Bur-nett and Pittman, 2005; Mazzucchelli et al., 2009). Recent findings indicate that ataxin-3 may

play a role in the organization of the cytoskeleton itself, since its absence leads to morpholog-ic alterations in cell lines, whmorpholog-ich are accompanied by the disorganization of the several cyto-skeleton constituents (microtubules, microfilaments and intermediate filaments) and a loss of cell adhesions (Rodrigues et al., 2010).

Adding to these functions, in vitro silencing of ataxin-3 in differentiating mouse my-oblasts, not only leads to the generation of an immature cytoskeleton, but also compromises myoblasts transition into muscle fibers, thus suggesting a potential role for ataxin-3 in myo-genesis (do Carmo Costa et al., 2010). This process comprises several events where remodel-ing of the cytoskeleton as well as of cell–cell and cell–extracellular matrix interactions is of crucial importance, with a tight control of protein expression and turnover (frequently by the UPP) being essential (Baylies and Michelson, 2001; Bryson-Richardson and Currie, 2008).

Page | 15

1.4.6. Nuclear inclusions and pathogenesis mechanisms

2005).

1.4.6.1. Nuclear inclusions

A common feature of polyglutamine diseases is the deposition of insoluble intracellular inclusions containing the misfolded disease protein (Paulson, 1999). Although their correlation with pathology is controversial (Bates, 2003; Michalik and Van Broeckhoven,

2003), nuclear and sometimes cytoplasmic inclusions are considered the identifying hallmark

of polyglutamine diseases (Ross and Poirier, 2004). The type of aggregated proteins and the regional and cellular distribution of the protein deposits vary from disease to disease, indicating that the individual proteins carrying the polyglutamine expansion modulate the associated pathogenesis (Paulson, 2003).

Mutant ataxin-3 in MJD, like other pathogenic proteins with expanded polyQ tracts, appears to undergo a conformational alteration and aggregate in cells forming ubiquitinated intranuclear neuronal inclusions (NIIs) with many different neurons containing more than one inclusion body, both in and outside areas affected by the selective neurodegeneration (Paulson et al., 1997b; Schimdt et al., 1998; Rub et al., 2006). Ataxin-3 aggregates were shown to contain -rich fibrillar structures of amyloid nature (Bevivino and Loll, 2001; Chen et al.,

2002). As for most amyloid-forming proteins, several pathways may drive the conversion of

the soluble protein to amyloid aggregates, through the formation of different conformationally altered monomeric or self-assembled multimeric species (Uversky, 2010). Polyglutamine monomers of ataxin-3 acquire β-strand conformations that have been shown to be cytotoxic in cultured cells (Nagai et al., 2007), assembling into oligomers (Bevivino and

Loll, 2001; Takahashi et al., 2008) able to simultaneously dissociate into monomers (Schaffar et al., 2004). Thus, it seems that β-stranded polyglutamine monomers are important for

pathogenesis in MJD and other polyglutamine diseases, however is still controversial if their toxicity is sufficient to be neurotoxic. Polyglutamine oligomers, have been shown to induce greater toxicity than polyglutamine monomers or inclusion bodies in neuronaly